For at least the last 2000 years, machair in the Outer Hebrides has undergone periods of
agricultural use (Ritchie 1967). Machair habitats have become fundamentally linked
with the crofting practices that take place on and around them. Machair grassland plant
communities have been influenced and maintained by rotational cultivation, grazing
practices and the use of kelp to trap moisture, fertilize soils and to stabilise the light
sandy surface soils typical of machair (Kent et al. 2003). Unfortunately, more intensive
agricultural practices are beginning to replace these traditional regimes, with over
grazing leading to reduced sward diversity in some areas and complete agricultural
abandonment and subsequent lack of grazing resulting in rank grassland vegetation, low
in plant species diversity, in other areas (Kent et al. 2003).
In conclusion, this study supports the existing body of evidence which suggests that the
restoration of grassland habitats is unlikely to be successful if it is to rely entirely on the
existing seed bank. This research has established that there is some correspondence
between the availability of a species in the above ground machair vegetation and its
presence in the seed bank. However, plant species differ in their persistence within the
seed bank and therefore, just because a species is present in the vegetation of machair
grassland does not mean we can assume that it will occur in the seed bank. This
suggests, in combination with the fact that key bumblebee forage plant species, typical
of machair grassland were absent from the seed bank, that additional seed input from
external sources is likely to be required in order to re-create wildflower-rich machair
habitat (Muller et al. 1998; Reiné et al. 2004). This will be particularly important if
machair grassland is to be restored in order to provide high quality foraging habitat for
C h a p t e r 5 | 109
the machair seed bank. Although the cost of implementing wildflower-rich seed mixes
can prove to be expensive (see chapter 6 of this thesis), the cost of restoration could be
reduced by disturbing the soil and adding only seed of those species such as C. nigra
C h a p t e r 6 | 110
6 Restoration and management of machair grassland for the
conservation of bumblebees
6.1 Abstract
Machair is a rare coastal grassland habitat, listed on the EU Habitats Directive, which
supports populations of nationally and internationally rare species including the
bumblebee species, Bombus distinguendus and Bombus muscorum. However, changes
in traditional land management practices have resulted in a loss of floral diversity in
some areas which has, in turn, reduced the availability of bumblebee foraging
resources. A restoration trial was established on a degraded machair site in western
Scotland and comprised four seed mixes and a fallow treatment which were monitored
over a three year period (2008-2010), in order to compare the relative abundance of
foraging bumblebees and the availability of forage plants. Two seed mixes contained
wildflower species identified as key bumblebee forage plants; one mix is currently used
to create bird and bee foraging habitat on nature reserves and the fourth mix is a
commercially available mix used for reseeding pasture. There was little variation in
forage availability and bumblebee abundance between treatments early on each year
(i.e. June 2008, 2009 and 2010) but marked differences emerged later in the season in
all three years. By the end of the monitoring period (August 2010), the two wildflower
treatments contained between four and eighteen times the number of inflorescences
than any other treatment type. Similar trends were observed in bumblebee abundance,
reflecting the availability of floral resources. Some of the rarest bumblebee species
exist primarily in restricted areas, which have largely escaped the intensification typical
of mainstream farming. In these areas it is important that habitat management is
C h a p t e r 6 | 111
suggest that the most effective method for restoring bumblebee forage plants on
machair is to sow wildflower-rich seed mixes which ensure the provision of forage
material throughout the season. The use of these mixes should be combined with late
cutting and winter grazing practices to maintain machair sward diversity over time.
6.2 Introduction
Machair is one of Europe‟s rarest habitats. It is limited in its global distribution to the
north and west of Scotland and Ireland and is listed on the EU Habitats Directive
(Angus 2001; Love 2003). Machair is described as a low-lying coastal grassland habitat
which forms on lime-rich soils comprising largely of blown sand, rich in shell derived
material (Ritchie 1967; Angus 2001; Love 2003).
Scottish machair supports nationally and internationally important populations of
several species. To date, research has largely focused on the importance of machair for
avian species, including corn bunting (Emberiza calandra), twite (Carduelis
flavirostris), corncrake (Crex crex), lapwing (Vanellus vanellus), redshank (Tringa totanus), dunlin (Calidris alpina) and ringed plover (Charadrius hiaticula) (Wilson
1978; Jackson & Green 2000; Wilson et al. 2007; Wilkinson & Wilson 2010).
However, a number of invertebrate species including the UK Biodiversity Action Plan
(UKBAP) priority species, the great yellow bumblebee (Bombus distinguendus) and the
moss carder bumblebee (Bombus muscorum), are also particularly associated with the
florally rich machair grassland of the Inner and Outer Hebridean islands (Benton 2006;
Goulson et al. 2006; Beaumont & Housden 2009).
Bumblebees have undergone substantial declines in recent decades and of the 25
C h a p t e r 6 | 112
species are severely threatened (Goulson 2003a). Agricultural intensification has been
held largely responsible for the decline of many species associated with farmland and
bumblebees are no exception (Chamberlain et al. 2000; Goulson 2003a; Goulson et al.
2008a).
The land management practices that have traditionally been implemented on machair
include low intensity grazing and rotational cropping and these crofting practices are
fundamentally linked to the floral diversity for which machair grassland habitats are
renowned (Roberts et al. 1959; Ritchie 1967; Owen et al. 2001; Kent et al. 2003).
However, the increasing modernisation, or conversely in some cases, the abandonment
of traditional crofting practices, has led to machair degradation (i.e. a decline in floral
diversity) in many areas (Angus 2001; Hansom & Angus 2005; Redpath et al. 2010).
This in turn is likely to have an impact on the species that currently thrive on machair
and in particular, the loss of florally rich machair swards poses a very real threat to
remaining populations of the UK‟s rarest bumblebee species.
This study examines the efficacy of five different machair restoration treatments in
providing foraging habitat for bumblebees on an area of degraded machair on the
southern Hebridean island of Oronsay. The principle aim of this research is to identify
the most effective treatment or treatments for restoring floral diversity to an
internationally rare habitat in order to provide resources for rare and declining Bombus
C h a p t e r 6 | 113
6.3 Methods
6.3.1 Study site
The island of Oronsay lies to the south west of the larger neighbouring island of
Colonsay, which is situated 32km west of the Scottish mainland. The area of machair
that was selected for this study had been heavily grazed by sheep for more than a
decade and consequently lacked floral diversity. A rotational cropping regime was
implemented at the site three years prior to the start of this experiment using oats, rye
and barley. This change in land management practice was an attempt to restore
traditional land management practices to the machair. However, like many crofts which
continue to be actively managed, inorganic fertilizer (NPK 16:10:10) was applied
annually at a rate of 500kg/ha, replacing the use of the traditional fertilisers such as
seaweed or farmyard manure.
6.3.2 Restoration treatments
In March 2007 the machair was ploughed and the ground prepared by adding a single
application of both agricultural lime and farmyard manure. Seaweed could not be used
on Oronsay as the island is a designated Special Protection Area (SPA) for red-billed
chough (Pyrrhocorax pyrrhocorax) and seaweed that is washed ashore is left on the
beaches to provide the choughs with suitable foraging habitat.
In May 2007 the site was divided in to 25 plots, each with a total area of 125m2 (5m x 25m). Five machair restoration treatments were implemented (table 6.1), each with five
replicates and the treatment plots were arranged in a quasi-complete Latin square design
so that each treatment type was adjacent to every other treatment type at least once.
C h a p t e r 6 | 114
would have been traditional on cultivated machair, a fast growing nurse crop of oats
and bere barley was sown over the treatments to protect the wildflower seedlings in the
early stages of growth and also to aid the stabilisation of the light, sandy soil (Roberts et
al. 1959). At the end of the first year (2007) the site was not cut or grazed in order to
allow the young plants to establish. In the second, third and fourth years of this study,
the treatments were cut and baled as silage in early September. Livestock in the form of
both cattle and sheep were put on to aftermath graze the site between September and
mid March. This system of rotational cutting and grazing is in accordance with
traditional crofting practice.
6.3.3 Vegetation surveys
The availability of bumblebee forage material in each of the 25 plots was monitored in
June, July and August over a three year period, from 2008 to 2010. Six 0.5mx 0.5m
quadrats were positioned at regular intervals in each plot, along a central transect line,
and the number of inflorescences of each bumblebee forage plant species present was
recorded. Bumblebee forage plant species were defined as the species known to be
utilised by bumblebees foraging on machair sites in western Scotland, as described by
Charman (2007) and Redpath et al. (2010).
6.3.4 Bumblebee surveys
In addition to monitoring the presence of forage plant species, each plot was also
surveyed for the presence of foraging bumblebees. Similarly to the vegetation surveys,
each plot was monitored in June, July and August throughout the three year period
(2008-2010). Each plot was surveyed for bumblebees twice, once in the morning and
once in the afternoon, and the total number of bees observed in each plot across the two
C h a p t e r 6 | 115
presence of foraging bumblebees in the first year (2007) in order to allow seedlings to
become established and to allow perennial species to flower.
A standard surveying methodology, adapted from Pollard (1977), was used to record
observations of foraging bumblebees. A transect was walked at a constant speed
through the centre of each plot, each walk taking an average of about five minutes. All
bumblebees observed foraging either side of the transect line, but within the plot area,
were recorded and identified to species level and all castes were combined. The plant
species on which bees were observed foraging were also recorded.
6.3.5 Data analysis
The availability of forage plant material (the number of inflorescences) in each
restoration treatment plot was examined using a generalised linear mixed effects model
(GLMM) with Poisson errors in the software package R version 2.10.1 (R Development
Core Team 2009). Treatment type, month and year were included in the model as fixed
factors and plot number was included as a random factor. All two- and three-way
interactions were explored and non-significant interactions were removed sequentially
using a backwards step-wise approach. Pair-wise post hoc comparisons were conducted,
using Tukey tests in the Multcomp package in R version 2.10.1, in order to assess
differences in forage availability among treatment types in each month and in each year.
The effect of machair restoration treatment on the abundance of foraging bumblebees
was also examined using a generalised linear mixed effects model with Poisson errors.
The availability of inflorescences was included in the model as an explanatory variable
and treatment type, month and year were included in the model as fixed factors. Plot
C h a p t e r 6 | 116
interactions were explored and non-significant interactions were removed sequentially
using a backwards step-wise approach. Pair-wise post hoc comparisons were
conducted, as before, using the Multcomp package in R, in order to assess the
differences in bumblebee abundances between treatment types in each month. A pseudo
R2 value (hereafter referred to as R2 values) was calculated for each GLMM, by correlating the values predicted by each model with the observed data (Zuur et al.
Table 6.1: Five machair restoration treatments and their definitions (prices accurate for 2010 and inclusive of VAT)
Treatment Type Cost of Seed (£/kg) Sowing Rate (kg/ha) Total Cost (£/ha) Definition
1. Wildflower 1 63.00 30.00 1890.00 Grass seed mixture containing plant species native to machair habitats but also known to be of importance for foraging bumblebees (ratio grass to wildflowers 80:20). These species were as follows: Lotus corniculatus, Arctium minus, Prunella vulgaris, Rhinanthus minor, Trifolium
pratense, Trifolium repens, Vicia cracca, Succisa pratensis, Thymus polytrichus, Cynosurus cristatus, Alopecurus pratensis, Festuca rubra ssp litoralis, Poa pratensis
2. Wildflower 2 200.00 20.00 4000.00 Wildflower 1, minus the grass species
3. Bird & Bee conservation
4.00 6.17 24.68 A brassica rich mix already implemented elsewhere in the Hebrides for the conservation of birds
but with added clover and phacelia to encourage foraging bumblebees. The mix contained kale, mustard, phacelia, fodder radish, linseed and red clover.
4. Commercial 4.50 29.65 133.43 A commercially available grass mix for re-seeding pasture. The mix contains approximately 5% white clover, an important bumblebee forage plant.
5. Fallow N/A N/A N/A The ground was ploughed and no seed was added. The vegetation was left to regenerate naturally from the existing seed bank.
C h a p t e r 6 | 118
6.4 Results